Model-based control
LQR and MPC
Students derive and discretize a state-space model, implement LQR and MPC, and compare their hardware behavior under real constraints.
Course overview
The course is organized around the full control engineering pipeline: modeling, sensing, embedded implementation, model-based control, observer design, reinforcement learning, deployment, and hybrid control.
Learning-based control
DQN, PPO, and TD3
Students train reinforcement learning policies in simulation and then study the sim-to-real gap during hardware deployment.
Integrated control
Hybrid and adaptive ideas
The final module combines switching-based hybrid control and residual model correction for improved robustness.